Asia’s Industrial Powerhouse Fuels Physical AI Adoption
The Asia-Pacific region, home to manufacturing powerhouses like South Korea, Japan, China, and Taiwan, is now driving the adoption of artificial intelligence in physical industries. Unlike economies dominated by services or software, these countries have long relied on large-scale production, export-driven industries, and highly optimized supply chains. This structural foundation is now shaping how AI is adopted and where investment flows, with companies like Config, a Seoul- and San Jose-based startup, securing backing from the venture arms of South Korea’s biggest manufacturers.
This mirrors what happened to the region’s semiconductor industry in the 2010s, when Asian companies like TSMC, Samsung, and SK Hynix became major players in the global market. Similarly, Asia’s manufacturing prowess is now fueling the growth of physical AI, with companies like Config providing the data layer for robotic foundation models (RFMs). Config’s oversubscribed $27 million seed round, led by Samsung Venture Investment, highlights the region’s growing interest in AI-powered robotics.
Config’s approach to providing data for robotics AI is gaining traction as large manufacturers increasingly seek to build their own proprietary robot AI instead of relying entirely on outside vendors. The startup’s role is similar to that of TSMC, which manufactures chips for Apple, Nvidia, and AMD without competing with any of them. Config aims to play a similar role in robotics by supplying the data, and its approach is resonating with investors and customers alike.
Config’s Decision Logic and Mechanics
Config’s decision to focus on providing data for robotics AI is driven by the high cost of gathering and labeling data for robot training. According to CEO Minjoon Seo, training large language models is expensive due to the computing power required, but the raw material – vast amounts of text from across the internet – is easy to obtain. In contrast, teaching robots to move requires physically collecting data, which is a costly and time-consuming process. Config’s approach is to provide a simpler and more cost-effective solution for companies looking to develop robot AI.
Config’s operational mechanics involve recording humans performing physical tasks in controlled studio environments and in the field. The startup operates out of Seoul and Hanoi, where a workforce of nearly 300 handles data production. To date, it has accumulated over 100,000 hours of human motion data, more than 30 times the size of AgiBot World, the largest comparable open-source dataset. Config’s approach to transforming the data before training begins is its core technical differentiator, and it has already generated revenue from its current customers, including large manufacturers, system integrators, and companies in the agriculture and defense sectors.
Config’s decision to focus on data production is also driven by the growing demand for proprietary robot AI among large manufacturers. According to Seo, companies are increasingly seeking to build their own robot AI instead of relying entirely on outside vendors. Config’s approach is to provide a cost-effective solution for these companies, and its role is similar to that of TSMC, which manufactures chips for major tech companies without competing with them.
Winners, Losers, and Disrupted Parties
Config’s approach to providing data for robotics AI is likely to benefit large manufacturers, system integrators, and companies in the agriculture and defense sectors. These companies will be able to develop their own proprietary robot AI using Config’s data, which will reduce their reliance on outside vendors and improve their competitiveness in the market. On the other hand, companies that specialize in robotics AI may be disrupted by Config’s approach, as they will face increased competition from large manufacturers that can develop their own robot AI using Config’s data.
Config’s approach is also likely to benefit the agriculture and defense sectors, which are major users of robotics AI. These sectors will be able to develop more advanced robotics AI using Config’s data, which will improve their productivity and competitiveness. On the other hand, companies that specialize in robotics AI for these sectors may be disrupted by Config’s approach, as they will face increased competition from large manufacturers that can develop their own robot AI using Config’s data.
The growth of physical AI in Asia is likely to be driven by companies like Config, which are providing the data layer for robotic foundation models (RFMs). This growth is likely to benefit the region’s manufacturing sector, which is a major user of robotics AI. On the other hand, companies that specialize in robotics AI may be disrupted by the growth of physical AI, as they will face increased competition from large manufacturers that can develop their own robot AI using Config’s data.
The Skeptical Case
One potential skeptical view of Config’s approach is that it may not be scalable. While Config has accumulated over 100,000 hours of human motion data, it is unclear whether this data can be easily replicated and scaled up to meet the needs of large manufacturers. Additionally, Config’s approach to transforming the data before training begins may not be effective for all types of robotics AI, and it is unclear whether the company’s approach can be adapted to different use cases.
Another potential skeptical view is that Config’s approach may not be competitive with established players in the robotics AI market. While Config has secured backing from the venture arms of South Korea’s biggest manufacturers, it is unclear whether the company can compete with established players like Physical Intelligence, Generalist AI, and Skild AI. Additionally, it is unclear whether Config’s approach can be differentiated from other companies that are providing data for robotics AI.
The Signal to Watch Next
One signal to watch next is Config’s ability to scale its data operation to one million hours of collected data. This will be a key indicator of the company’s ability to meet the needs of large manufacturers and to compete with established players in the robotics AI market. Additionally, it will be important to watch Config’s progress in launching its cloud-based Robot-as-a-Service product, which will allow companies to run Config’s foundation model without requiring onboard hardware.
Another signal to watch next is the growth of physical AI in Asia. This growth is likely to be driven by companies like Config, which are providing the data layer for robotic foundation models (RFMs). It will be important to watch the adoption of physical AI in different sectors, including manufacturing, agriculture, and defense, and to monitor the impact of this growth on companies that specialize in robotics AI.
Pick one tactic from this post and apply it today. Which one will you start with?
By Daniel Cross, Digital Growth Strategist at TrendFlashy
Ready to launch your own asset?
Check out our guide on Building a Profitable Online Business.
